Smartphone-based digital markers and clinical symptoms during therapy for Borderline Personality Disorder

No established digital markers yet exist for borderline personality disorder (BPD) or for predicting symptom change throughout therapy. This study adopts an exploratory approach to uncover candidate associations between smartphone behavior and BPD symptoms.We objectively recorded smartphone use in individuals with BPD (N = 49) during an eight-week inpatient DBT program. Symptom severity was assessed using the Borderline Symptom List (BSL-23) pre- and post-DBT, and the Dissociative Experiences Scale (DES-II) at baseline. Smartphone usage time and communication behavior (phone calls, SMS) were compared with t-tests to student samples. We explored Spearman correlations between smartphone usage and pre- and post-therapy BSL, DES-II as well as BSL symptom change. To predict treatment outcomes by smartphone variables, we modeled a multiple linear regression controlling for age, baseline symptom severity, and pretreatments.Compared to student samples, individuals with BPD used their smartphones more than twice as long (5.88 h per day) and made a higher number of phone calls. Higher average smartphone usage time was associated with a greater trait tendency to dissociate (DES-II), whereas no other digital markers were linked to baseline BPD symptoms. Locked screen time was linked to poorer symptom improvement, while a higher number of incoming calls was associated with better treatment outcomes. These effects already emerged in the first week smartphone use and explained 60% of the variance.Excessive smartphone use and call behavior may function as digital markers of BPD, with more dissociative patients showing particularly high use. Higher locked screen time, potentially reflecting maladaptive checking behavior, was associated with poorer treatment outcomes, whereas a greater number of incoming calls, possibly indicative of social support, was associated with better improvement. These findings suggest that digital phenotyping may help to identify early markers of recovery when integrated into psychotherapy. However, due to the small sample size, the unmatched student control group, and the exploratory nature of the analyses, the results of the study should be interpreted with caution, and their generalizability should be examined in future studies.

retrospectively registered - https://osf.io/dfq9y/?view_only=4c19b891bb6448009b22f60b2552bd73.